Radworks Seed Node (RSN) Tokenomics Considerations

steakhouse · May 22, 2024, 7:02pm

We’re building on the RSN Org proposal posted a few weeks ago and wanted to share some thoughts on tokenomics for Radicle:

Foundations of DePIN

DePIN has become an increasingly important part of the wider crypto ecosystem and its mindshare. According to Kaito, in 2024, the category has risen from negligible levels to >2% of the entire market’s mindshare.

Source: Kaito

The most successful DePIN network to date is Bitcoin. By providing programmatic incentives to the miners, the Bitcoin network has achieved an impressive hash rate of 640 EH/s. This level of computing power is equivalent to that of 6.4 billion Nvidia RTX 4090s!

Source: Blockchain.com

The follow-on DePIN projects have followed the foundational principles of Bitcoin’s success to bootstrap the supply-side of their respective network. The basic components are:

A large incentives pool which follows either a time/block-based or a KPI-based emission schedule. Inspired by Bitcoin, many DePIN projects’ emissions follow some variation of a decay function
Proportional distribution of rewards: Rewards are distributed proportional to some quantitative measure - hash rate in Bitcoin, stake in Ethereum, storage power in Filecoin. In the case of Saturn, rewards are based on a composite performance score (e.g. Time to First Byte, Bandwidth Served, Download Speed, Node Uptime)
Permissionless competition: Open the network up to anyone to participate such that the level of service provision naturally improves as the network becomes more competitive

Various DePIN projects to date have made different design choices when implementing the three pillars. For instance, when determining the incentives pool, projects decide on parameters such as how much to incentivise, in what currency (token or fiat), and whether the emissions will be formulaic or governance-adjustable.

When determining the proportional distribution of rewards, the considerations are which behaviors to incentivize or penalize through slashing, if staking is enabled. The extent of the openness of the network could also vary. While Bitcoin and Ethereum have allowed permissionless entry from the beginning, Lido has taken a more progressive route to decentralization (e.g. a curated node operator module) in an effort to ensure a higher quality service in the bootstrapping stage of the network build out. Render is another DePIN project that is selectively onboarding supply.

In reality, the impact of the design choices that projects make to bootstrap a new economy and the associated trade-offs, are not very well understood. This is because many DePIN projects are still in their relatively early stages of experimentation. Even those that have somewhat successfully scaled their supply-side, such as Filecoin and Helium, have yet to achieve a sustainable level of demand.

DePIN mechanism design

A comprehensive study undertaken by 1kx has explored the tokenomics design choices of DePIN projects. Projects can be generally categorized along four dimensions:

Limited vs. unlimited supply: Limited supply has the benefit of enforcing a notion of token scarcity, but has the downside that the network should be successfully bootstrapped to reach a sustainable state before maximum supply is reached. From the projects’ perspective, fixed supply imposes a level of capital discipline (the do more with less mindset), but also could limit the flexibility necessary for growth. Bitcoin’s fixed supply tokenomics has set an odd precedent for the follow-on crypto networks
Formulaic vs. governance-adjustable: A formulaic approach (e.g. BTC block rewards halving every 210k blocks) may gain social acceptance and an “immutability premium”, at the expense of inflexibility down the road. This approach places much pressure to get the tokenomics right at the start with incomplete understanding of a complex system. Conversely, adaptive methods through governance-adjustable mechanisms offer flexibility to make corrections over time as more information becomes available,though they may risk diminishing stakeholder trust if voting power is concentrated among the founding team and early investors.
Constant vs. decaying rewards: Are the amount of emissions chosen constant or deflationary based on certain parameters, such as time or performance?
Time/KPI-driven: Should emissions be purely based on time/blocks or depend on the state of the network such as node operator KPIs?

Source: 1kx

The above classification reveals two interesting trends.

DePIN projects, likely inspired by Bitcoin, often follow a decaying token emission schedule that compensates early network participants more generously. The rewards decrease as the network matures
However, unlike Bitcoin, most projects employ DAO governance for flexibility, allowing for adjustments but risking stakeholder acceptance with significant changes

The study also shows monthly token inflation by emissions categories. Somewhat intuitively, the fixed, decaying model has the highest front-loaded inflation. Most DePIN projects reach peak token inflation in the first two years.

The Bitcoin-inspired model, while most followed by DePIN projects, have resulted in a few issues. In the next section, we investigate the outcomes from earlier projects and attempt to infer useful learning points or lessons.

Source: 1kx

Case Study: Filecoin

Filecoin is an OG DePIN project which has iterated on Bitcoin’s minting model. It employs hybrid minting, consisting of “Simple Minting”(30%) and “Baseline Minting” (70%).

Simple Minting is the Bitcoin-like exponential decay, whereby 330m FIL tokens (out of max supply of nearly 2 billion FIL ) are released on a 6 year half-life based on time. This means 97% of tokens will be released over the next 30 years.
Baseline Minting is KPI-based emissions. 770M FIL tokens are minted based on the capacity of the network. Baseline minting operates by comparing the network’s actual storage capacity to a predefined “baseline” storage capacity curve. If the network’s storage capacity meets or exceeds the baseline, new FIL tokens are minted at an optimal rate. If the actual storage is below the baseline, the minting rate decreases. All tokens will be released if the Filecoin network reaches a Yottabyte of storage capacity in under 20 years.

Filecoin’s hybrid emissions is arguably a step forward vs. simple minting, as baseline minting in theory should lead to increased creation of value for the network and lower risk of minting FIL too quickly. However, we observe two main weaknesses with the model:

Baseline minting still does not take into account the demand side of the network. During 1kx’s data sample period between October 2020 to April 2023, nearly 280M FIL (equivalent to $9.5 billion using monthly average FIL prices!) were distributed to the supply side. Lagging the time by a year, cumulative protocol revenues from October 2021 to April 2024 were 11.2M FIL (or $166m). The implied return on this “customer acquisition spend” was a mere 4.0% in FIL, or even worse at 1.7% in dollar terms. By not considering the demand response to the supply build out, Filecoin has so far shown poor returns on its incentivisation “investment.”
According to researchers from CryptoEconLab, accounting for demand in the tokenomics design has proven to be difficult due to a lack of reliable ways of reducing Sybil attacks. For instance, if KPIs related to data stored were to feature in the incentive design for storage providers (SPs), the optimal move could be to fake data storage in order to qualify for higher incentives.
Parameterisation of simple minting half-life. Part of the over-incentivisation that we’ve seen may be due to the fact that the 6 year half-life may have been too long. Juan Benet of protocol labs has said that if he could do the tokenomics again, he would shorten the half-life from 6 years to 2 years.

From the above, we take away early learnings about token-led DePIN economies:

Study your demand: To create a sustainable economy, it is crucial to balance demand growth with token inflation used for supply growth. Can we anticipate who the users of the service offered by the network will be? Can we pace the rate of emissions to balance supply growth with demand growth better?
There are risks with over-incentivisation: Highly favorable reward system could do more harm than good in drawing mercenary capital early on. Filecoin’s supply side (indicated by the green line, Network Raw Byte Power) grew rapidly and peaked in 2022. Since then, supply has been on a downward trajectory and has been undershooting the long-term baseline power target. Admittedly, FIL price cratering has not helped the attractiveness of being a storage provider. But one wonders to what extent, the earlier storage providers joined purely for incentive farming

Source: Starboard

FIL price (Source: Coinmarketcap)

Study the cost base of the supply side: One practical solution that we see being effective for curbing overincentivsation and mercenary capital is designing incentives around the estimated cost base of the supply side.

Suppose a storage provider (SP) in Europe has an estimated cost base of $x, a SP in China has an estimated cost base of $0.8x, and a SP in America $1.4x. Could the reward program mint some amount of native token at each time interval that will cover the highest cost base supplier (i.e. $1.4x)?

Consider rewarding the supply side in fiat pricing rather than in native token: A tokenomics researcher, Uroš Kalabić, who has peer reviewed the Helium’s Burn-and-Mint Equilibrium (BME) tokenomics goes a step further and suggests that the solution to the avoiding overincentivisation that risks crashing the network economy is simply by incentivising the supply side in stablecoins rather than the native token.

He describes the Burn-and-Mint-Equilibrium as a Goldilocks problem that needs to offer an optimal inflationary balance between the acceptability by the token holders and the supply side contributors. If the supply-side incentives are paid in the native token, the network runs into two possible scenarios:

The purple line (undermint): Supply-side contributors do not receive enough incentives which decays in an exponential manner (right) and therefore the value of the network flatlines (left).
The blue line (overmint): Conversely, the system provides too much to the contributors (right), that eventually the token holders abandon the network due to hyperinflation and the value goes to 0 (left)

Source: Uroš Kalabić

With some math, Kalabić concludes that the solution to the overmint/undermint Goldilocks problem is to price the mint rewards to contributors in fiat too.

The Radicle Seed Network (RSN)

A simplified value flow of the Radicle Seed Network (RSN) look as follows:

The tokenomics group to consider the following aspects:

i) Understand supply-side cost base

Define the hardware requirements
Model the cost of running nodes
Would the onboarding be permissionless or curated?

ii) Define KPIs to incentivise

What metrics should be used to monitor “good performance”? (e.g. Bandwidth, Download Speed, Node Uptime).
How do we easily verify these metrics?
What is the right weighting formula between the KPIs?

iii) Define system penalties

Do we enable staking and slashing? What are the pros and cons?
Slashing criteria

iv) Understand demand side potential

Who is likely to use this service? What would the user pay for such a service?
What % of Radicle users would be interested?

v) Define the incentive pool and emissions schedule

How much should be reserved for RSN incentives?
Systematic vs. Governance-adjustable
Decay vs. Fixed
Fiat pricing vs. Token pricing

General Recommendations

Tokens are a scarce resource - use them strategically and sparingly!
- Steakhouse played a pivotal role in preserving Lido DAO’s resources by recommending the discontinuation of $LDO incentives to LPs. Following our proposal, the DAO ceased these payouts without experiencing any loss in liquidity.
Think about token incentives in the context of CAC (Customer Acquisition Cost). From the DAO’s perspective, we want the lowest CAC to successfully bootstrap a network that efficiently equilibrates the supply of node operators/service providers and demand

Source: Dune Analytics (@steakhouse)

b3nj4min · May 22, 2024, 9:49pm

Dear Radicle DAO,

Thank you for sharing your insightful tokenomics considerations for the Radicle Seed Network (RSN).

I commend the thorough analysis of the DePIN landscape and the case study on Filecoin’s hybrid minting model. Your findings highlight crucial lessons for designing sustainable token economies, particularly the importance of balancing supply-side incentives with demand growth and understanding the cost base of node operators.

After carefully reviewing your proposed value flow for RSN and the key aspects to consider, I would like to provide some recommendations and ideas to help optimize the tokenomics design:

Supply-side cost base:
- Conduct a comprehensive study on the hardware requirements and associated costs for running RSN nodes across different regions. This will help determine the optimal level of incentives needed to attract and retain high-quality node operators.
- Consider a curated onboarding process for node operators in the initial stages to ensure a reliable and efficient network. As the network matures, gradually move towards a permissionless model to encourage decentralization and competition.
KPI-based incentives:
- Define clear and measurable KPIs that align with the network’s objectives, such as bandwidth, download speed, and node uptime. Ensure these metrics are easily verifiable to prevent Sybil attacks and maintain the integrity of the incentive system.
- Develop a weighted formula for KPIs that prioritizes the most critical factors for user experience and network performance. Regularly review and adjust these weights based on user feedback and network data.
System penalties:
- Implement a staking and slashing mechanism to discourage malicious behavior and ensure node operators have “skin in the game.” However, be cautious not to set the stakes too high, as it may deter participation from smaller players and lead to centralization.
- Define clear and fair slashing criteria that penalize severe infractions while allowing room for honest mistakes and temporary downtime.
Demand-side potential:
- Conduct market research to identify the target users for RSN and their willingness to pay for decentralized storage and retrieval services. This will help gauge the potential demand and inform the token emissions schedule.
- Consider offering a freemium model or trial period to attract users and showcase the benefits of RSN. Monitor the conversion rates from free to paid users to refine the pricing strategy.
Incentive pool and emissions schedule:
- Be strategic and conservative in allocating tokens for RSN incentives. Regularly assess the network’s growth and adjust the incentive pool accordingly to avoid over-inflation and ensure long-term sustainability.
- Opt for a governance-adjustable emissions schedule that allows for flexibility in responding to changing market conditions and network needs. Start with a moderately decaying emissions curve and adjust based on demand growth and user feedback.
- Explore the feasibility of using stablecoins or a basket of currencies for pricing RSN incentives to mitigate the volatility associated with the native token. This can help attract a broader range of node operators and users.

As you refine the tokenomics design, prioritize mechanisms that minimize the risk of over-incentivization and mercenary behavior. Learn from the experiences of other DePIN projects and adapt their best practices to suit RSN’s unique requirements.

Furthermore, I strongly advise framing token incentives as a customer acquisition cost (CAC) from the DAO’s perspective. By continuously monitoring and optimizing the CAC, you can ensure the most efficient use of the DAO’s resources in bootstrapping a thriving ecosystem of node operators and users.

Lastly, maintain open communication channels with the community and be transparent about the tokenomics design process. Regularly seek feedback and be open to iterating on the model as the network evolves.

I believe that by carefully considering these recommendations and ideas, Radicle DAO can develop a robust and sustainable tokenomics model for RSN that drives long-term growth and value creation. I look forward to seeing the progress and success of this ambitious endeavor

miohtama · May 23, 2024, 8:16am

Focusing too much on the supply side and pumping the token too much is not productive or serves only short-term goals. Demand is much more important. Most of the blockchain projects are ghost towns with zero demand, so we get all kinds of token engineering to have artificial short-term pumps.

Also, any tokenomics can be changed afterwards: see Aave, Yearn, Ethereum, etc.

There are some exceptions, though. In the Web3 social media space, Farcaster, and Lens to some degree, have shown that there can be a revenue stream before a token.

Thus, the most important question is likely to be:

Who is likely to use this service? What would the user pay for such a service?

There has not yet been a decentralised application that would beat their centralised counterparty, maybe with the exceptions of Uniswap and Aave. Thus, seeing how to get Githubuseres, the de facto market leader (with 90% marketshare?) to switch will be hard.

Pro users are happy with Github free tier
Enterprise users won’t touch token

This is especially hard, since many of these Github users hate the idea of cryptocurrencies and tokens, as we see many times in e.g. Hacker News discussions about Radicle

(You can find many other posts on the similar topic).

It’s a difficult question where is the product market-fit. One idea could would be trying to convince large cryptocurrency projects to switch from Github to Radicle and have high annual fees, but this would be sweet-pilled by having some kind of token swaps or accepting their own token in Radicle ecosystem somehow.

This is because

Crypto foundations are rich
Crypto foundations are pro-crypto
Crypto foundations want to have more utility for their tokens

shelb_ee · May 27, 2024, 9:50am

Thank you @steakhouse / @equanimiti for the post! This provides helpful context on specific considerations for the RSN and the community while thinking about tokenomics approaches.

I would be curious to hear @lftherios 's reaction to the points made above and how he thinks they might be used/referenced by the tokenomics working group.

tokenarchitecture · June 4, 2024, 5:48am

Sharing our comments below from Just Cryptoeconomics:

Supply-Side Participation

Open Participation vs. Curated Onboarding:

There is currently no definitive evidence favouring either fully opening up supply-side participation (e.g., Bitcoin, Ethereum) or utilising a degree of curation (e.g., Lido, Render) for Decentralised Physical Infrastructure Networks (DePIN) systems.
If opting for a curated onboarding process, the associated costs must be accounted for. Modelling these costs should include a buffer to prevent pricing out operators in more expensive markets and centralising node operators geographically.

Hardware Requirements:

It is crucial to understand how far into the future hardware requirements can be predicted and how subject to change these requirements are.

Key Performance Indicators (KPIs)

Framework for Weighting KPIs:

Develop a framework for determining the appropriate weighting formula for KPIs.
Consider if this framework is realistic or if it differs too much between projects.
A lightweight mental framework might be helpful, though this could be overly simplistic.

Token Incentives and Costs

Token Incentives and Customer Acquisition Cost (CAC):

We agree that token incentives should be considered in the context of CAC.

Opportunity Cost:

Evaluate the opportunity cost for supply-side providers who might choose more profitable alternatives.
Assess if resources are mutually exclusive between projects and if providers must choose one over another.
If the market overcompensates providers relative to demand, determine the appropriate response.

System Penalties

Contextualisation of System Penalties:

Provide more context on defining system penalties.
Identify projects that employ slashing conditions and explore the relationship between the strictness of these conditions and the expected rewards.

Stablecoins and Hybrid Solutions

Tradeoffs of Using Stablecoins:

Consider the tradeoffs of using stablecoins and explore hybrid solutions.
For example, in the Filecoin case, minting could initially use FIL tokens, with KPI-based emissions in stablecoins, creating a buy-back and distribute model.

Formulaic vs. Governance-Adjustable Models

Combination of Formulaic and Governance-Adjustable Models:

Explore implementations that combine formulaic and governance-adjustable distribution methods.
Outside of DePIN, Cosmos Hub is an example of a network adjusting its formulaic distribution through governance.

Fixed Supply Regimes

Bitcoin’s Precedent:

Bitcoin’s fixed supply regime has set a unique precedent for subsequent crypto networks.
Projects like Helium have adopted a similar philosophy but with modifications such as diminishing emissions (e.g., HIP-20). This approach imposes time pressure to find alternative incentivisation methods as emissions decrease, potentially requiring demand-side taxation or altering supply dynamics.